AI System Aims to Prevent Tapir Deaths on Brazilian Roads

In Brazil, a computer scientist named Gabriel Souto Ferrante has developed an AI-based solution to detect and warn drivers about the presence of certain medium and large-sized animal species, such as tapirs, on roads to prevent animal deaths and accidents. Ferrante created a database of thousands of images of these animals and trained an AI model to recognize them in real-time.

The project, which has been published in the journal Scientific Reports, aims to collaborate with road management companies to access traffic cameras and relay real-time warnings to drivers, similar to navigation apps. This technology is intended to reduce wildlife strikes and save human lives, as previous strategies like signage and animal crossing structures have proven insufficient in addressing the scale of the problem.

Why this matters: Brazil's Brazilian Center for Road Ecology estimates that around 475 million vertebrate animals die on the country's roads every year. This AI system has the potential to significantly reduce roadkill incidents, protect wildlife, and improve road safety for drivers in Brazil and beyond.

The AI system uses a network of cameras equipped with software that can recognize various animal species with high precision. When an animal is detected near the road, the AI triggers an alert system to warn drivers, either through roadside signs or integration with GPS navigation apps.

Researchers are also exploring the use of YOLO (You Only Look Once) deep learning models for real-time wildlife detection on roads, with the Scaled-YoloV4 model showing promising results in minimizing false negatives and the nano version of YoloV5 achieving the fastest processing speed.

The project's developer, Gabriel Souto Ferrante, hopes to reactivate a previous app called Urubu to help identify roadkill hotspots and raise public awareness. "This technology aims to reduce wildlife strikes and save human lives, as other strategies such as warning signs and animal crossing structures have proven insufficient to address the scale of the problem," Ferrante stated.

The inventive AI method to roadkill prevention is currently awaiting support from road management companies and funding to be fully implemented. Challenges such as infrastructure integration and reliable internet connectivity need to be addressed before the technology can be widely deployed across Brazil's road network. However, if successful, this innovative approach could serve as a model for other countries facing similar issues with wildlife-vehicle collisions.

Key Takeaways

• Brazilian scientist developed AI to detect animals on roads and alert drivers.
• AI system uses cameras and deep learning models to recognize wildlife in real-time.
• Aims to reduce roadkill incidents, protect wildlife, and improve road safety in Brazil.
• Awaiting support from road management companies and funding for full implementation.
• Innovative approach could serve as a model for other countries with similar issues.